Mohamed Oubbati

About

Mohamed Oubbati is a researcher whose work sits at the intersection of neural computation, adaptive control, and autonomous robotics. His research focuses primarily on applying recurrent and reservoir computing architectures — most notably Echo State Networks (ESNs) — to real-world robot control problems, alongside broader investigations into reinforcement learning, neurocontrol, and multi-agent coordination. Among his most significant contributions is the integration of Echo State Networks within adaptive critic and actor-critic frameworks, enabling mobile robots to anticipate future rewards and learn optimal control policies online — a technically challenging problem in continuous time and space. His 2005 work on kinematic and dynamic adaptive control of nonholonomic robots using recurrent neural networks, alongside his research on omnidirectional RoboCup players, demonstrated that recurrent architectures can deliver robust, real-time motion control in demanding physical environments. His neural field approaches to multi-robot formation control further showcased his breadth, offering elegant solutions to obstacle avoidance and geometric coordination simultaneously. With citations accumulating across robotics, machine learning, and neural systems communities, Oubbati's body of work reflects a sustained commitment to understanding how embodied agents can learn adaptive, intelligent behavior through principled neural architectures — a perspective increasingly central to modern AI research.

Research Focus

Key Achievements

Top Papers

1. 1
   Adaptive Critic Design with Echo State Network

   23 citations · 2010
2. 2
   Kinematic and Dynamic Adaptive Control of a Nonholonomic Mobile Robot using a RNN

   23 citations · 2005
3. 3
   Velocity Control of an Omnidirectional RoboCup Player with Recurrent Neural Networks

   17 citations · 2006
4. 4
   A neural framework for adaptive robot control

   15 citations · 2009
5. 5
   Neural Fields for Controlling Formation of Multiple Robots

   8 citations · 2007
6. 6
   Anticipating Rewards in Continuous Time and Space with Echo State Networks and Actor-Critic Design.

   7 citations · 2011
7. 7
   Learning Robot-Environment Interaction Using Echo State Networks

   4 citations · 2010
8. 8
   Learning of embodied interaction dynamics with recurrent neural networks: some exploratory experiments

   4 citations · 2014
9. 9
   NEURAL FIELDS FOR BEHAVIOR-BASED CONTROL OF MOBILE ROBOTS

   4 citations · 2006
10. 10
    Dynamic Task Assignment in a Team of Agents

    4 citations · 2006

Key Collaborators

Contact & Links